Our main vaccine platform is based on recombinant vesicular stomatitis virus (rVSVs), a live-attenuate vaccine approach. Over the years we have generated several rVSVs expressing the glycoproteins (GP) of representative isolates of all species of Ebola virus: Sudan ebolavirus (SEBOV), Zaire ebolavirus (ZEBOV), Tai forest ebolavirus (TFEBOV), Bundibugyo ebolavirus (BEBOV) and Reston ebolavirus (REBOV). Additionally, we generated rVSVs expressing the GPs of two isolates of Marburg virus: Lake Victoria marburgvirus isolate Musoke (MARVmus) and Angola (MARVang). All vaccine vectors have been extensively characterized in cell culture and their protective efficacy has been evaluated in animal models against homologous challenges including nonhuman primates. In an effort to decipher the mechanism of protection of the rVSV vaccine vectors we used the rVSV/ZEBOVgp as a model. We could demonstrate in nonhuman primates that antibodies specific to the viral antigen play a critical role in protection. Ongoing work is investigating the mechanism of protection of the rVSV vaccine vector against MARV. Defining the mechanism(s) and correlate(s) of protection will be milestones for moving the rVSV platform into clinical trials. Cross-protection among the different Ebola virus species and even Marburg virus is an important consideration, but is thought to be difficult to achieve due to relatively high genetic and antigenic variability among genera in particular, but also among species within a single genus, and the general lack of cross-protective antibodies even among species. In a first attempt to address this issue we previously used a single-injection protocol with three blended vaccine vectors (rVSV/SEBOVgp, rVSV/ZEBOVgp and rVSV/MARVmusGP) and demonstrated complete protection against challenge with the three homologous virus species. We have also performed another proof-of-concept study, in which we evaluated cross-protection following immunization with a single vaccine vector (rVSV/ZEBOVgp or rVSV/CIEBOVgp) and demonstrated partial cross-protection against challenge with a heterologous virus species (BEBOV). This demonstrates that monovalent rVSV-based vaccines may be useful against a newly emerging species; however, heterologous protection across species remains challenging and may depend on enhancing the immune responses either through booster immunizations or through the inclusion of multiple immunogens. Overall, we can conclude that single monovalent rVSV vaccine vectors can provide partial cross-protection in cases of challenge virus species that are genetically more closely related. As mentioned above, one approach to overcome this limitation is the use of blended monovalent rVSV vaccine vectors, which provide broader protection against homologous and partial protection against certain heterologous challenges. Another approach to overcome the limitations in cross-protection is the use of multivalent rVSV vaccine vectors. In a proof-of-concept study protection against ZEBOV and Andes virus (ANDV) challenge was demonstrated using a single rVSV vector expressing both the ZEBOVgp and ANDV glycoprotein in the Syrian hamster model. This data showed that bivalent rVSV vectors are a feasible approach to vaccination against multiple pathogens. Further, this study demonstrated that the Syrian hamster is an adequate model to study rVSV-mediated protection prior to nonhuman primate work. Based on the results described above, we have in the past fiscal year generated more bivalent and trivalent rVSV vectors expressing two or three different filovirus GPs, one as a transmembrane protein (replacing the VSV glycoprotein) and one or two as soluble proteins that will be secreted during vector replication. Recovery of these recombinant vaccine viruses turned out to be difficult and in vitro characterization of the already rescued vectors is ongoing. Efficacy testing of these vectors will be performed initially using rodent models, mainly the Syrian hamster. The most promising vaccine vectors will be moved into efficacy testing in the nonhuman primate model for filovirus infections.